00:00.58 | cr2 | dcordes: have you picked the .py files too ? |
00:01.34 | dcordes | cr2: I fetched both haret.exe and console from the urls Kevin2 posted |
00:01.52 | cr2 | 94e00080 should be parsed |
00:02.28 | cr2 | ok, the .py is not updated |
00:02.35 | cr2 | change it to 0x80 and 0x84 |
00:02.43 | dcordes | k |
00:03.29 | dcordes | 007.099 IRQS IRQ1: IRQ33(33)=1 |
00:03.29 | dcordes | 007.099 IRQS IRQ0: IRQ0(0)=1 IRQ1(1)=1 |
00:03.29 | dcordes | 007.099 IRQS IRQ1: IRQ32(32)=1 IRQ33(33)=1 |
00:04.56 | cr2 | 32 and 33 are gpio bank1 and 2 mux |
00:05.04 | Kevin2 | tmzt_: http://sourceforge.net/project/showfiles.php?group_id=173455&package_id=198682 |
00:05.40 | Kevin2 | cr2: I don't think it's 0x80/0x84 on non-A variant. It would be very odd to get two irqs in the same event. |
00:06.27 | dcordes | cr2: http://linuxtogo.org/~lgorris/haretlog-20080928_020304.log |
00:06.33 | cr2 | Kevin2: yes. it looks strange |
00:06.43 | cr2 | Kevin2: then we need to separate A and non-A |
00:07.20 | cr2 | dcordes: where is the patch by dzo ? |
00:07.35 | cr2 | dcordes: he must handle the irqs somehow :) |
00:08.02 | dcordes | cr2: you mean the vogue kernel? |
00:08.05 | cr2 | yes |
00:08.29 | dcordes | http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=shortlog;h=refs/heads/htc-vogue |
00:10.32 | cr2 | hm. |
00:10.39 | cr2 | the irq.c is from A |
00:11.52 | Kevin2 | cr2: There is a vogueirq.c file. |
00:12.00 | dcordes | yep |
00:12.45 | dcordes | cr2: check the commits from 4 days ago to see what's vogue specific |
00:12.51 | dcordes | 9 days |
00:12.56 | cr2 | yes, he uses 0x0 and 0x4 |
00:13.22 | dcordes | http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=commit;h=3cc470413d630ead203bc045ba5d81576ae189f1 |
00:13.51 | cr2 | <PROTECTED> |
00:13.51 | cr2 | <PROTECTED> |
00:14.03 | cr2 | so we need separate versions |
00:16.13 | *** join/#htc-linux sikkle (n=sikkle@modemcable132.57-201-24.mc.videotron.ca) |
00:16.49 | Kevin2 | Will ram start always be 0x20000000 on A variant? |
00:17.06 | cr2 | Kevin2: there are 2 banks |
00:17.41 | cr2 | Kevin2: raphael has 128M@0x10000000 and 128M@0x20000000 |
00:18.17 | Kevin2 | The patch NetRipper sent had "set ramaddr 0x20000000" for Raphael. |
00:18.22 | cr2 | Kevin2: the alloctest shows that the RAM is allocated from 0x2* area first and then goes down to 0x1* |
00:19.15 | cr2 | SanMehat said msm kernel does not support discontiguous memory |
00:19.54 | cr2 | so we are limited to 1 bank |
00:20.20 | cr2 | one can only guess how G1 has 192MB. in the first bank ? strange. |
00:28.35 | Kevin2 | cr2 / dcordes: Can you try http://www.handhelds.org/~koconnor/haret/haret-20080927b.exe and http://www.handhelds.org/~koconnor/haret/haretconsole-20080927b.tar.gz ? |
00:30.16 | cr2 | ok |
00:37.37 | cr2 | Kevin2: works for me(tm) |
00:38.10 | cr2 | the gpio irq decoder need proper masks, but i'll do it tomorrow (aka today) |
00:38.37 | Kevin2 | cr2: Ok - I'll check in what I have. |
00:39.13 | cr2 | dcordes: already sleeping ? |
00:41.28 | Kevin2 | cr2: If you have the irq names you can program them into haretconsole too - take a look at how pxa does it. |
00:42.17 | cr2 | yes. |
00:42.28 | *** join/#htc-linux dcordes_ (n=dcordes@unaffiliated/dcordes) |
00:42.54 | cr2 | dcordes_: the new haret works for me. can you test it ? |
00:46.25 | dcordes_ | cr2: want the log? |
00:46.36 | cr2 | does it work ? |
00:46.47 | dcordes_ | what exactly? |
00:46.58 | cr2 | it seems that raphael uses the builtin touchscreen controller |
00:47.02 | cr2 | tracing |
00:47.20 | dcordes_ | the spam soup seems similar to the one before |
00:47.26 | Kevin2 | Hrmm. I forgot to set msm7xxx back to 0x00 |
00:48.04 | Kevin2 | dcordes_ try http://www.handhelds.org/~koconnor/haret/haret-20080927c.exe |
00:53.27 | dcordes_ | Detected machine Generic MSM7xxx/MSM7xxx (Plat='PocketPC' OEM='HTC_Kaiser') |
00:54.16 | Kevin2 | How does "wirq" look? |
00:54.17 | dcordes_ | now it only shows irq 33 |
00:56.46 | Kevin2 | dcordes_: So, is it the same behavior prior version of haret had? |
00:57.26 | dcordes_ | yes |
00:58.24 | Kevin2 | Ok - I've committed the changes. |
00:58.42 | dcordes_ | nice |
00:58.47 | dcordes_ | cr2: do you know what kaiser irq 33 is? |
00:58.57 | dcordes_ | cr2: 0x21 33 INT_GPIO_GROUP2 |
00:59.57 | cr2 | dcordes_: it's the bank2 gpio |
01:00.37 | cr2 | the status register needs to be added |
01:00.54 | dcordes_ | a bank gpio? |
01:00.57 | cr2 | but it is necessary to calculate the masks first |
01:01.06 | cr2 | no, some gpio from bank2 |
01:01.24 | cr2 | to decode which one exactly it's necessary to do some minor work |
01:01.33 | cr2 | heh. |
01:01.52 | cr2 | i alwaysforget that on non-pxa 'wi' does not do mmutrace |
01:03.12 | Kevin2 | mmutrace works on non-pxa. |
01:03.46 | cr2 | with wi ? |
01:03.55 | Kevin2 | Yes. |
01:04.10 | Kevin2 | (How else would one use mmutrace?) |
01:04.24 | cr2 | yes ... i don't see anything... |
01:04.42 | Kevin2 | Make sure you have all the virtual mappings covered. |
01:06.51 | cr2 | yes |
01:06.56 | dcordes_ | cr2: any hint on the acx base addres? |
01:07.04 | cr2 | 0x9* is created by p2v() |
01:07.17 | cr2 | but the "real" address is 0xb* |
01:07.57 | cr2 | yes, i see i2c |
01:08.12 | Kevin2 | cr2: p2v can be misleading with mmutrace. It's safer to find all the mappings via "dump mmu 2 <adddr>" and then program them into mmutrace. |
01:08.31 | cr2 | that's true. |
01:08.46 | cr2 | but i think it's in the docs :) |
01:10.31 | dcordes_ | :( |
01:13.50 | cr2 | aha. |
01:13.56 | cr2 | adc is used. |
01:17.02 | cr2 | in module: touch.dll (78780000 - 78791000) |
01:19.13 | cr2 | ok, enough for today. good night. |
01:20.57 | Kevin2 | cr2: Good night. |
01:37.51 | bryan__ | Is there a way i can use Haret to probe my device to see where my sd card is located .. im getting the SD card not found error ... im assuming my sd card isnt at /dev/block/mmcblk0p1 as this init wants it to be . |
01:40.02 | dcordes_ | bryan__: try /dev/mmcblk0p1 ? |
01:40.26 | bryan__ | ok, |
01:46.01 | tmzt_ | bryan__: what kernel (zImage) and initrd are you using? /dev/block/mmcblk0p1 is a feature of devfs which has not been used in recent kernels |
01:47.00 | bryan__ | the android for vogue , figured vogue is the best choise for my device . both with msm7500 |
01:47.19 | bryan__ | its the only one that boots and gives me any error too . |
01:47.52 | bryan__ | there a better version i should be using ? |
01:48.21 | tmzt_ | do what dcordes said, then, where did you get the instructions to use /dev/block/mmcblk0p1? |
01:49.03 | bryan__ | its just what was in there . |
01:49.43 | bryan__ | ive just pulled files off the net .. im not much of a coder . |
01:49.44 | tmzt_ | so it boots on iris? |
01:49.53 | bryan__ | ya it trys |
01:50.06 | tmzt_ | how far does it go? |
01:52.04 | bryan__ | not very far at all .. i get Welcome to vogue linux then it starts with the sd error .. and continues with m5m_i2c errors . cant get bus? |
01:52.31 | tmzt_ | you get white text on a black screen "Welcome to vouge linux" ? |
01:52.43 | bryan__ | ya ... pengiun and all . |
01:52.44 | tmzt_ | is it msm_i2c? |
01:52.53 | bryan__ | ya prolly .. small text |
01:53.39 | tmzt_ | you device is cdma? |
01:54.01 | bryan__ | ya. |
01:54.27 | bryan__ | that giving the error ? |
01:54.47 | tmzt_ | kernel boots, but the i2c is different, you will proably have to probe with haret from wince to find out where they are. |
01:55.18 | bryan__ | ill have to ten rebuild the kernal from there ? |
01:55.39 | tmzt_ | do you get a detected line in haret before booting, or are you using startup.txt ? |
01:56.39 | bryan__ | startup.txt |
01:56.58 | tmzt_ | can you rename it to default.txt ? |
01:57.15 | tmzt_ | http://wiki.xda-developers.com/index.php?pagename=HTC_Iris |
01:57.20 | tmzt_ | that one? |
01:57.41 | bryan__ | thats it .. |
01:57.52 | bryan__ | i cant see the detect screen . |
01:58.30 | tmzt_ | not in the text box at the bottom? |
01:58.37 | tmzt_ | it's qvga same as my moto q |
02:00.01 | bryan__ | im sorry the detect line is ..? |
02:00.12 | bryan__ | screen |
02:00.41 | tmzt_ | can you make move the focus to that box, press the ok button, and scroll (I forget) |
02:02.40 | bryan__ | i can see the launch script box and the info about connecting on port 9999 i can scroll to select the listen option i assume it is but i cant see anything lower then that. |
02:03.13 | tmzt_ | ok |
02:04.04 | bryan__ | im gonna have to jet in a about ten min .. but ill be back online around 10:45ish |
02:04.57 | tmzt_ | I guess you need ginge, dzo, or cr2 to help you, if you want to work on this. It's msm7500 but I guess it's different from vouge/titan |
02:06.05 | bryan__ | i should use haret to prob the device tho .. that is my next step ? |
02:06.26 | tmzt_ | you need some way of connecting it to your computer and getting an ip address, like activesync or rndis, and the haretconsole from http://handhelds.org/~koconnor/haret/ |
02:06.31 | tmzt_ | which needs python |
02:06.43 | bryan__ | yes i was reading about that . |
02:06.49 | tmzt_ | you are using the haret that came from the bootkit? |
02:07.05 | bryan__ | no .. the newest one |
02:07.11 | tmzt_ | which one? |
02:07.24 | bryan__ | either from the 25th .. or there was one i saw posted in here earlier . |
02:07.54 | tmzt_ | Kevin2: which is the right haret for msm7500? |
02:09.13 | bryan__ | should i still try /dev/mmcblk0p1 in the init file |
02:09.18 | bryan__ | ? |
02:10.13 | tmzt_ | yes, you have something extracted on your sd card? |
02:10.41 | tmzt_ | and you copied haret and kernel over to storage on your iris? |
02:10.59 | bryan__ | no . |
02:11.21 | bryan__ | well i have all on sd. |
02:12.00 | bryan__ | so system and data on the sd and init zimage on iris . |
02:12.07 | bryan__ | with haret . |
02:12.23 | tmzt_ | what is the command line in your default.txt? |
02:13.03 | bryan__ | set cmdline "ppp.username=mobile@jamamobile ppp.password=telecom msmsdcc_fmax=48000000 quiet" |
02:14.12 | tmzt_ | where does the /dev/mmcblk0p1 go then? |
02:14.30 | bryan__ | thats in my init . |
02:14.44 | bryan__ | i can put that in the cmdline ? |
02:15.00 | tmzt_ | in what init, you mean in the android image? |
02:15.21 | tmzt_ | what is your initrd line? |
02:15.53 | bryan__ | set initrd initrd.gz |
02:16.05 | bryan__ | which ive un gz to just initrd |
02:17.08 | bryan__ | ill be back in a bit .. |
02:17.25 | bryan__ | :) |
02:18.20 | *** join/#htc-linux dcordes (n=dcordes@unaffiliated/dcordes) |
02:28.57 | Kevin2 | tmzt_: I'd use the latest cvs build - I have a copy at: http://www.handhelds.org/~koconnor/haret/haret-20080927c.exe and http://www.handhelds.org/~koconnor/haret/haretconsole-20080927b.tar.gz |
02:30.57 | tmzt_ | ok, I think he left, but bryan__ is trying to make his htc iris work with the vogue kernel, and gets an error about msm_i2c not finding the bus, but linux boots |
02:31.21 | tmzt_ | would that be because the addresses changed and how can he find out? |
02:41.14 | Kevin2 | Usually, one can't use a kernel that was built for a different machine. |
02:42.37 | Kevin2 | It's a fairly involved process to map out all the devices present on a phone. I'd use mmutrace on haret to see if/how wince accesses each device. |
02:51.50 | *** join/#htc-linux ALoGeNo (n=QUAKEIII@245.Red-81-44-9.dynamicIP.rima-tde.net) |
02:52.20 | ALoGeNo | hi |
03:08.14 | ALoGeNo | everybody sleeping? |
03:08.14 | ALoGeNo | :) |
03:14.24 | tmzt_ | no, what do you need? |
03:19.27 | ALoGeNo | hey tmzt_ =D |
03:20.08 | ALoGeNo | no nothing rigth now im just reading about uni in handhelds org :) |
03:20.34 | ALoGeNo | im rolling with cvs stuff :) |
03:21.04 | ALoGeNo | ill go try compile new kernel for the uni |
03:22.28 | tmzt_ | did you get the initfs from MnemonicPunk ? |
03:24.31 | ALoGeNo | hum.. nope tmzt_ i was two day without sleep the other time sorry i cant remember |
03:25.05 | ALoGeNo | im downloading kernel sources from handhelds dot org |
03:25.49 | ALoGeNo | tmzt_: are you talking abaout the img file of the debian one? |
03:26.49 | ALoGeNo | sorry but i mix things about 900c and uni when talk with you heh |
03:29.35 | ALoGeNo | tmzt_: i have 900c stoped with the message..kernel panic - not syncing: Attempted to kill init! |
03:29.47 | ALoGeNo | jlime kernel and debian :) |
03:29.56 | ALoGeNo | hda2 detected |
03:30.12 | ALoGeNo | VFS mounted ext2 |
03:33.12 | ALoGeNo | tmzt_: unable to open initial console, but in default.txt i have /bin/bash |
03:34.00 | ALoGeNo | maybe i need to set the permissions in /bin under linux or what? |
03:41.19 | tmzt_ | set CMDLINE "root=/dev/hda2" |
03:41.25 | tmzt_ | that's all I have |
03:41.37 | ALoGeNo | okey ill test :) |
03:41.47 | ALoGeNo | thnak you :) |
03:41.59 | tmzt_ | how did you install debian? |
03:42.03 | ALoGeNo | tankui ;) |
03:42.08 | ALoGeNo | yeeep |
03:42.15 | ALoGeNo | bootstraped |
03:42.21 | bryan__ | heya |
03:42.31 | ALoGeNo | hey bryan__ |
03:42.52 | bryan__ | hey hey . |
03:43.23 | bryan__ | so am i actually getting anywhere with this vogue kernal ? |
03:43.57 | ALoGeNo | damm i really need to make me a kitchen with something to overwrite the gx.dll, o let that overwrited becose im bored of overwrite that crap xd |
03:45.27 | tmzt_ | just make an empty file gx.dll in the same directory |
03:46.24 | ALoGeNo | yep i copy and paste this file when i reboot.., |
03:47.03 | tmzt_ | you can put it in your fat partition with haret |
03:47.16 | Kevin2 | bryan__: You're leaps ahead by virtue of having linux to successfully boot with lcd support working. It's still quite a bit of work to get the rest going though. |
03:47.30 | Kevin2 | ALoGeNo: What's wrong with gx.dll on your device? |
03:47.42 | ALoGeNo | yep is do in that way thanks tmzt_ :) |
03:48.18 | ALoGeNo | same crap of message.. |
03:48.37 | ALoGeNo | unable to open initial console |
03:48.55 | tmzt_ | we are using cmonex wince image on mp900c, which has a broken gx.dll so haret can't find VRAM the usual way |
03:49.01 | ALoGeNo | the FS is ok.. but it cant run the bash.,. |
03:49.15 | tmzt_ | how did you install debian? |
03:49.29 | ALoGeNo | yep i need do that.. |
03:49.44 | Kevin2 | tmzt_: If gx.dll isn't present, then haret does find the vram? |
03:50.03 | ALoGeNo | i was follow that... letme copy.. |
03:50.22 | ALoGeNo | debootstrap --include=wireless-tools,hotplug --arch arm --foreign etch /etch http://ftp.uk.debian.org/debian |
03:50.23 | tmzt_ | I think haret doesn't start at all |
03:50.43 | bryan__ | kevin2: i should go thru the steps of mmutrace as you stated. that is my next place to start ? |
03:50.54 | tmzt_ | ALoGeNo: you need to finish the installation |
03:50.55 | ALoGeNo | later copy in the card and run kernel with haert |
03:50.59 | ALoGeNo | haret |
03:51.12 | Kevin2 | tmzt_: Haret doesn't require gx.dll - it will try to pull it in if it's found. It only uses it if GETRAWFRAMEBUFFER method doesn't work. |
03:51.31 | tmzt_ | this is on ce.net 4.2 |
03:51.33 | ALoGeNo | how?, putting command in default.txt? |
03:52.20 | ALoGeNo | tmzt_: i know that i have to do that debootstrap --second-stage , but i cant in to the bash shell |
03:52.32 | Kevin2 | bryan__: Yeah. You need to document all devices present on your device. |
03:53.04 | Kevin2 | tmzt_: I don't understand. Haret crashes if an empty gx.dll isn't in the same directory as haret? |
03:53.18 | ALoGeNo | no |
03:53.43 | ALoGeNo | haret crashes if the fake gx.dll isnt the /windows |
03:53.58 | ALoGeNo | in the* |
03:54.07 | tmzt_ | ALoGeNo: what does it say when it crashes? |
03:54.17 | bryan__ | will do :) |
03:54.45 | ALoGeNo | when i boot haret with thw original gx.dll in /windows haret hang the machine |
03:55.19 | ALoGeNo | if i put and overwrite the file gx.dll with faked one haret boots normally |
03:55.31 | Kevin2 | ALoGeNo: Create a file called "earlyharetlog.txt" in the same directory as haret.exe. Then see what is in "haretlog.txt" after you try to start haret. |
03:55.52 | ALoGeNo | tnx Kevin2 |
03:56.06 | Kevin2 | ALoGeNo: I'm curious to see what's in haretlog.txt - can you paste it somewhere for the successful and failed cases? |
03:57.25 | ALoGeNo | what the f#king reset button isnt just in keyborad? hehe |
03:57.40 | tmzt_ | there's a button on the bottom |
03:57.55 | ALoGeNo | yes.. sorry. |
03:58.01 | ALoGeNo | why* the f#king reset button isnt just in keyborad? hehe |
03:58.03 | ALoGeNo | xd |
04:02.46 | ALoGeNo | haretlog is empty Kevin2 |
04:03.31 | Kevin2 | But the file exists? |
04:03.40 | Kevin2 | Can you post the log after a successful boot? |
04:04.15 | ALoGeNo | nope.. i booted the device with haretlog.txt inside and is empty let my try again |
04:04.38 | Kevin2 | You must create the file "earlyharetlog.txt" -- the resulting log is in "haretlog.txt". |
04:05.49 | ALoGeNo | ok |
04:07.56 | ALoGeNo | booting |
04:10.29 | ALoGeNo | unable to open initial console |
04:11.05 | ALoGeNo | this is the default.txt.. |
04:11.06 | ALoGeNo | set KERNEL zImage |
04:11.07 | ALoGeNo | #set INITRD initrd.gz |
04:11.07 | ALoGeNo | set MTYPE 659 |
04:11.07 | ALoGeNo | set CMDLINE "console=tty0 console=ttyS0,115200n8 rootdelay=5 root=/dev/hda2" |
04:11.07 | ALoGeNo | boot2 |
04:13.11 | ALoGeNo | this the haret Kevin2 http://rafb.net/p/rgKqd434.html |
04:13.19 | ALoGeNo | haretlog* |
04:13.59 | Kevin2 | ALoGeNo: Can you post the haretlog.txt? |
04:14.10 | Kevin2 | Oops - old message. |
04:14.15 | ALoGeNo | yep this link http://rafb.net/p/rgKqd434.html |
04:14.22 | ALoGeNo | np |
04:14.22 | ALoGeNo | :) |
04:14.40 | ALoGeNo | paste is for free gratis cero nada :D |
04:16.45 | Kevin2 | ALoGeNo: You have a really old haret. Can you grab the latest from: http://www.handhelds.org/~koconnor/haret/haret-20080927c.exe |
04:17.18 | ALoGeNo | ohh thank you very much Kevin2 :) |
04:18.05 | Kevin2 | Let me know if the new one ever fails to start for you. |
04:18.36 | ALoGeNo | well i get that one becose that is for the mp900 iget it form jlime dot com |
04:19.56 | Kevin2 | Haret isn't device specific anymore. |
04:20.15 | ALoGeNo | okey ill try |
04:24.03 | ALoGeNo | okey np with new haret but "Warning: unable to open an initial console" |
04:24.48 | ALoGeNo | hey that isnt need a swap file or something? |
04:25.10 | ALoGeNo | in jlime in that step mount the swap and later boots |
04:27.35 | tmzt_ | ALoGeNo: you don't have /dev/console, which is okay if udev is configured, but you haven't finished --second-stage yet |
04:28.05 | ALoGeNo | okey what i need to do tmzt_? |
04:28.32 | ALoGeNo | mkedir /dev/console? |
04:28.43 | ALoGeNo | mkdir /dev/console?* |
04:29.22 | tmzt_ | mknod /dev/console c 5 1 |
04:31.12 | ALoGeNo | tmzt_: but that under my laptop, inside the folder that i have mounted the car fs no? |
04:32.05 | tmzt_ | yes |
04:32.17 | tmzt_ | mknod /card/dev/console c 5 1 |
04:34.19 | ALoGeNo | okee dokee |
04:39.23 | ALoGeNo | looks like wants.. |
04:39.30 | ALoGeNo | mmmmand.... |
04:40.00 | ALoGeNo | is taken a while |
04:41.46 | tmzt_ | what's taking a while? |
04:42.31 | ALoGeNo | is stoped in Freeing init memory: 88K |
04:42.40 | ALoGeNo | _ |
04:43.09 | ALoGeNo | now is when it creates the swap.. |
04:43.10 | tmzt_ | it still sees hda2 and you get a VFS message? |
04:43.18 | ALoGeNo | yep |
04:43.37 | ALoGeNo | the file system is mounted ok |
04:43.55 | ALoGeNo | mounted root ext2 filesystem |
04:45.36 | ALoGeNo | if i insert the pcmcia that response.. and moount it |
04:45.37 | ALoGeNo | ;) |
04:46.14 | ALoGeNo | and 33V 33V |
04:46.32 | tmzt_ | you dont't have the cf card in when you boot? |
04:47.26 | ALoGeNo | nope that is necesary? |
04:47.51 | tmzt_ | the card with your root filesystem on it |
04:48.03 | ALoGeNo | oh yep of course |
04:48.20 | ALoGeNo | i say the wifi card |
04:49.16 | tmzt_ | ok, add init=/bin/sh to CMDLINE |
04:49.54 | ALoGeNo | ok |
04:53.48 | ALoGeNo | the same.. tmzt_ |
04:54.37 | ALoGeNo | i dont need to create the swap filesystem? |
04:59.24 | *** join/#htc-linux PoohbaLT1 (n=Poohba@c-98-235-52-97.hsd1.nj.comcast.net) |
05:03.58 | tmzt_ | only if you run out of memory |
05:06.49 | ALoGeNo | that is waiting in Freeing init memory: 88K |
05:07.04 | ALoGeNo | and nothing happens |
05:09.02 | ALoGeNo | tmzt_: you tald me something about eabi really? |
05:09.11 | ALoGeNo | for compile kernel maybe? |
05:10.49 | tmzt_ | you get "Freeing init memory" and nothing else |
05:11.05 | ALoGeNo | yep |
05:11.15 | ALoGeNo | atop |
05:11.19 | ALoGeNo | stop* |
05:11.30 | ALoGeNo | but isnt hanged |
05:15.21 | tmzt_ | you have init=/bin/sh ? |
05:18.13 | ALoGeNo | yep i try that way and init=/bin/bash too |
05:28.38 | tmzt_ | if you type, the characters echo? |
05:29.50 | ALoGeNo | unope keyboard dont work |
05:29.52 | bryan__ | does the watchgpio command not work ? |
05:30.41 | ALoGeNo | i dunno |
05:42.18 | tmzt_ | bryan__: what are you asking, we are talking about after the kernel has booted? |
05:43.00 | tmzt_ | oh, the documentation is outdated, I think you need wi 20 |
05:43.32 | bryan__ | im connected to the device thu telnet. |
05:44.54 | bryan__ | trying to use commands like DUMP GPIOST |
05:45.15 | bryan__ | WATCHGPIO but they dont work |
05:46.58 | Kevin2 | bryan__: Those commands are old. Follow the documentation at: http://www.handhelds.org/moin/moin.cgi/HaRET_20Documentation |
05:48.37 | *** join/#htc-linux LunohoD_ (n=alex@e180069134.adsl.alicedsl.de) |
05:50.00 | bryan__ | ahh yes, thank you. Those are the docs i couldn't find ..... |
06:14.26 | *** join/#htc-linux ALo (n=QUAKEIII@26.Red-81-37-222.dynamicIP.rima-tde.net) |
06:27.06 | tmzt_ | ALoGeNo: what's in your default.txt file? |
06:35.08 | *** join/#htc-linux rmoravcik (n=rmoravci@ip-89-102-255-171.karneval.cz) |
06:45.00 | *** join/#htc-linux goxboxlive (n=goxboxli@139.84-48-212.nextgentel.com) |
07:21.58 | *** join/#htc-linux cousin_luigi (n=luigi@unaffiliated/cousinluigi/x-395723) |
07:22.00 | cousin_luigi | hello |
07:22.20 | cousin_luigi | is the htc tytn significantly faster than the s720? |
07:30.45 | *** join/#htc-linux lkcl__ (n=lkcl@87.112.82.199.plusnet.ptn-ag2.dyn.plus.net) |
07:36.25 | *** join/#htc-linux nebi (n=nebi@c-498fe253.02-145-7570701.cust.bredbandsbolaget.se) |
07:37.00 | *** join/#htc-linux nebi_ (n=nebi@c-498fe253.02-145-7570701.cust.bredbandsbolaget.se) |
07:43.44 | *** join/#htc-linux LunohoD_ (n=alex@e180072046.adsl.alicedsl.de) |
07:44.01 | *** join/#htc-linux nebi_ (n=nebi@c-498fe253.02-145-7570701.cust.bredbandsbolaget.se) |
07:49.50 | *** join/#htc-linux sikkle (n=sikkle@modemcable132.57-201-24.mc.videotron.ca) |
08:03.33 | *** join/#htc-linux ex87 (n=ex87@78.84.228.138) |
08:03.53 | *** join/#htc-linux pleemans (n=toi@d54C2AAB7.access.telenet.be) |
08:27.50 | *** join/#htc-linux Tellmedark (i=dcfd4205@gateway/web/ajax/mibbit.com/x-cfc43d3bf25d43e8) |
08:30.48 | *** join/#htc-linux olesis (n=olesis@ool-4351a418.dyn.optonline.net) |
08:30.51 | olesis | hi |
08:34.15 | Tellmedark | hi |
08:34.38 | paulproteus | Howdy Tellmedark. |
08:48.31 | *** join/#htc-linux Zoolooc (n=fredsiba@nrbg-4dbfc515.pool.einsundeins.de) |
09:01.28 | *** join/#htc-linux dzo_ (n=dzo@121.98.128.127) |
09:07.30 | *** join/#htc-linux LunohoD (n=alex@e180072098.adsl.alicedsl.de) |
09:16.18 | *** join/#htc-linux lkcl_ (n=lkcl@87.112.82.199.plusnet.ptn-ag2.dyn.plus.net) |
09:21.53 | *** part/#htc-linux cousin_luigi (n=luigi@unaffiliated/cousinluigi/x-395723) |
09:39.06 | *** join/#htc-linux lkcl__ (n=lkcl@87.112.82.199.plusnet.ptn-ag2.dyn.plus.net) |
09:57.35 | *** join/#htc-linux MnemonicPunk (n=atlantis@dslb-084-060-033-089.pools.arcor-ip.net) |
10:00.45 | MnemonicPunk | Hello. =) Does anyony know which oe branch I should use when I want to build something for the Magician? |
10:02.09 | *** part/#htc-linux Tellmedark (i=dcfd4205@gateway/web/ajax/mibbit.com/x-cfc43d3bf25d43e8) |
10:08.07 | *** join/#htc-linux nebi (n=nebi@c-498fe253.02-145-7570701.cust.bredbandsbolaget.se) |
10:11.50 | *** join/#htc-linux miknix (n=miknix@bl5-240-155.dsl.telepac.pt) |
10:32.52 | *** join/#htc-linux lpotter (n=ljp@CPE-124-191-144-181.vic.bigpond.net.au) |
10:54.10 | *** join/#htc-linux diogene31_ (n=rj@mur31-2-82-243-122-54.fbx.proxad.net) |
11:03.18 | cr2 | hi |
11:03.40 | *** join/#htc-linux Miek (n=mike@unaffiliated/mikechml) |
11:26.48 | *** join/#htc-linux Dinde (n=kayser@sur-internet.net) |
11:35.25 | *** join/#htc-linux exco (n=exco@e181072093.adsl.alicedsl.de) |
11:47.28 | *** part/#htc-linux rmoravcik (n=rmoravci@ip-89-102-255-171.karneval.cz) |
11:49.57 | *** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
11:50.03 | *** join/#htc-linux rmoravcik (n=rmoravci@ip-89-102-255-171.karneval.cz) |
12:13.25 | NetRipper | hi :) |
12:16.41 | cr2 | hi NetRipper |
12:16.59 | cr2 | NetRipper: raph100 uses builtin touchscreen |
12:17.29 | NetRipper | sounds logical |
12:17.29 | NetRipper | ;) |
12:17.46 | NetRipper | external touchscreen would be unhandy |
12:19.10 | cr2 | vogue uses tsc2003 i2c chip |
12:20.30 | NetRipper | by builtin you mean built into cpu? |
12:20.47 | cr2 | yes |
12:22.24 | *** join/#htc-linux SNAKE-AFK (n=s@dyndsl-085-016-202-129.ewe-ip-backbone.de) |
12:23.06 | NetRipper | OK |
12:23.09 | NetRipper | -caps |
12:23.45 | NetRipper | i'd hope android kernel has drivers for it |
12:25.23 | NetRipper | regarding the framebuffer stuff, i talked with kevin2 yesterday and he indicated the irqs should really be disabled after the take_control() but that Sleep() probably re-enables them. and the reason the framebuffer isn't drawn properly, is that msm has optimizations that make lcd not redraw directly, but it needs a signal via a reg or similar |
12:26.45 | cr2 | yes. so you can forget about using framebuffer |
12:27.12 | cr2 | but you may try to set the flashlight gpio :) |
12:27.18 | NetRipper | well i may try making the fb code better |
12:27.25 | NetRipper | to signal the lcd to redraw |
12:27.53 | cr2 | it may be too difficult |
12:28.29 | cr2 | android code needs to be rechecked 10000 times |
12:28.33 | cr2 | <PROTECTED> |
12:28.33 | cr2 | <PROTECTED> |
12:28.34 | cr2 | <PROTECTED> |
12:28.34 | cr2 | <PROTECTED> |
12:28.34 | cr2 | <PROTECTED> |
12:28.36 | cr2 | <PROTECTED> |
12:28.51 | cr2 | i trace this data without 0x100 offset |
12:29.41 | cr2 | i didn't find the builtin touchscreen driver in android git. does it exist ? |
12:30.59 | NetRipper | don't know |
12:31.06 | NetRipper | maybe not yet there |
12:31.44 | cr2 | addlist mmutrace 0xb2300000 4 |
12:31.55 | cr2 | addlist mmutrace 0xb230000c 4 |
12:32.04 | cr2 | this traces i2c reads and writes |
12:32.56 | cr2 | capacitive touchscreen :) |
12:33.07 | NetRipper | mm |
12:33.58 | cr2 | touchscreen registers are documented in wiki |
12:34.35 | NetRipper | what's the haret commadn to trace that? |
12:35.15 | cr2 | wi N |
12:35.29 | cr2 | N in seconds |
12:40.55 | cr2 | hehe. there is a hidden 'fieldtest.exe' app |
12:42.05 | cr2 | it just parses the at@ data. |
12:42.20 | *** join/#htc-linux Dinde (n=kayser@sur-internet.net) |
12:43.23 | cr2 | i'd like to know what a bright microsoft head decided to put 2000+ files into windows directory ;-) |
12:43.56 | NetRipper | lol |
12:44.06 | NetRipper | forgot to put p2v() around irq addresses |
12:44.15 | NetRipper | hangs haret ;p |
12:44.36 | cr2 | you don't need them here |
12:44.45 | cr2 | they are virtual |
12:44.50 | cr2 | 0xb23* |
12:44.57 | NetRipper | yea |
12:44.59 | NetRipper | but for addlist irqs |
12:45.04 | cr2 | p2v() returns 0x932* |
12:45.08 | cr2 | which will not work |
12:45.13 | cr2 | ok |
12:47.24 | cr2 | it's fun that htc put their own dlls into \windows directly, so you can just copy them with the file manager :) |
12:47.59 | NetRipper | hehe yes |
12:48.08 | NetRipper | back on hermes they put it on XIP |
12:48.25 | NetRipper | or at least htccamera dll was |
12:48.26 | NetRipper | ;) |
12:50.00 | NetRipper | hoop onderzoekjes opeens |
12:50.01 | NetRipper | ;) |
12:50.19 | NetRipper | er, wrong chan |
12:51.50 | tcccp | neuken keuken? |
12:51.55 | swetland | cr2: G1 doesn't use the onboard touchpanel control |
12:52.20 | swetland | cr2: so we haven't spent much time mucking with it |
13:03.20 | *** join/#htc-linux Dinde896 (n=kayser@81-65-130-197.rev.numericable.fr) |
13:08.52 | cr2 | swetland: thanks. it does not look very difficult to write a driver for it. |
13:11.50 | cr2 | swetland: what is AD5 on msm ? |
13:16.32 | NetRipper | a 'word' is 4 bytes, right? |
13:20.35 | cr2 | 2 bytes ? |
13:20.57 | cr2 | but i'm an oldtimer :) |
13:28.57 | NetRipper | hehe |
13:29.42 | cr2 | the i2c clock is at 0x68 though |
13:30.15 | NetRipper | arf i cant get gpios |
13:30.41 | NetRipper | im trying to set gpio58 (flashlight with question mark) |
13:30.56 | NetRipper | but when i write the value, and then read it again, nothing changed |
13:31.45 | cr2 | read , | , write to out ? |
13:31.57 | NetRipper | eh |
13:32.00 | NetRipper | write to out? |
13:32.02 | cr2 | you may check how it's done in android kernel |
13:32.04 | cr2 | yes |
13:32.11 | NetRipper | im using the in2 |
13:32.24 | NetRipper | thought you had to write to in |
13:32.43 | NetRipper | so, read from in2, and write to out2 |
13:32.46 | NetRipper | let's try that |
13:33.10 | cr2 | in is for in, out is for out |
13:33.20 | NetRipper | yea ok but you can also look from the side of the cpu |
13:33.28 | NetRipper | you set something to 'in' for input of cpu |
13:33.35 | NetRipper | my twisted world |
13:34.38 | *** join/#htc-linux Poohba1 (n=Poohba@c-71-58-20-66.hsd1.nj.comcast.net) |
13:34.44 | cr2 | i#m busy with i2c/dex |
13:35.29 | cr2 | btw, you can calculate the bank masks from http://wiki.xda-developers.com/index.php?pagename=Raphael_GPIO, so we will add them to .py trace decoder |
13:36.32 | cr2 | it's 0xffff for 1, |
13:37.24 | NetRipper | ohhh lovely i got my first gpio to work |
13:37.34 | NetRipper | flashlight is confirmed then btw, so i'll remove that question mark |
13:37.34 | cr2 | 0x7fffff for bank2 |
13:37.45 | cr2 | softreset |
13:38.02 | cr2 | some people burnt the led in universal's time ;) |
13:38.17 | NetRipper | yes i put back the original value |
13:38.23 | NetRipper | that i read, without the flashlight bit |
13:38.25 | NetRipper | and it turned off |
13:38.44 | cr2 | question was not that seriously meant :) |
13:39.14 | NetRipper | and i heard about the burned led |
13:39.21 | NetRipper | which is why i made a time limit in torchbutton app |
13:39.45 | cr2 | hm, 0xffff for 0, 0x7fffff for 1 |
13:39.48 | cr2 | ok |
13:39.52 | NetRipper | but torchbutton uses camera dll instead of gpios |
13:40.04 | cr2 | it's more portable |
13:40.33 | cr2 | htc was very nice. |
13:41.00 | cr2 | i see "ldr" at the address which haret shows :) |
13:42.22 | cr2 | 0x1ffffff for 2 |
13:42.24 | cr2 | err |
13:42.43 | cr2 | 0x7ffffff for 1 |
13:42.48 | NetRipper | what are those masks for? |
13:42.56 | NetRipper | they prevent tracing some gpios? |
13:43.10 | cr2 | 0x7ffffff for 3 |
13:43.10 | NetRipper | like spamming ones |
13:43.19 | cr2 | yes, the top bits |
13:43.44 | cr2 | which are reserved |
13:43.51 | NetRipper | aha |
13:44.01 | cr2 | spamming can be masked manually with 'ibit gpios X' |
13:44.26 | cr2 | 0xfff for 4 |
13:45.00 | cr2 | 0x7fff for 5 |
13:46.42 | cr2 | so, 7201A : 0xffff for 0, 0x7ffffff for 1, 0x1ffffff for 2, 0x7ffffff for 3, 0xfff for 4, 0x7fff for 5 |
13:49.00 | cr2 | heh. dzo should not edit android drivers directly ;-) |
13:51.40 | *** join/#htc-linux ex87 (n=ex87@78.84.228.138) |
13:52.06 | NetRipper | im trying to see where to set the masks |
13:52.21 | cr2 | wait a bit. |
13:57.55 | cr2 | to the 'statN' here http://wiki.xda-developers.com/index.php?pagename=RaphaelMemoryMap |
14:03.32 | ex87 | Hello! Please show me right direction to get cross-compiler for TyTN II kernel compilation (For Kubuntu 8.04 box). |
14:07.18 | cr2 | ex87: i think this one -> wget http://qtopia.net/downloads/neo/arm920t-eabi-ficgta01-toolchain.tgz |
14:07.47 | cr2 | at least i use it |
14:08.45 | NetRipper | cr2, but you mentioned adding the masks to the .py files |
14:08.56 | cr2 | NetRipper: doing it now. |
14:08.58 | NetRipper | cr2, don't see a place to do that |
14:11.20 | cr2 | i need to calculate the gpio offsets |
14:11.37 | swetland | cr2: AD5 -> ADSP5 "application dsp" |
14:13.08 | cr2 | swetland: ok. |
14:15.06 | ex87 | cr2, thanks, will try |
14:17.03 | dcordes | ex87: there are two problems: the sd driver is not 100% ok, so sometimes errors occur when more files on the card are accessed simulateneaously and the framebuffer driver does not allow X to work |
14:17.28 | dcordes | so it's almost pointless to insteall a distro like debian/ubuntu on the kaiser |
14:17.55 | dcordes | if you do a very tiny install it might boot |
14:19.57 | Kevin2 | cr2: If you build up a gpio mask via ibit calls, then you can read out the full mask by running "show gpios". |
14:21.16 | dcordes | cr2: acx? |
14:22.00 | ex87 | dcordes: Main reason why I want to compile kernel is for testing android and checking out what's wrong with tsc2003.c (touchscreen calibration). I'm newbie in all this stuff, but want to learn, so I choosed to check out htc-linux |
14:22.53 | dcordes | ex87: ok. htc-linux doesn't have the latest code. it is better to checkout android-msm-2.6.25 from the google tree |
14:23.12 | cr2 | Kevin2: i' masking the undefined gpio irq lines |
14:24.05 | dcordes | ex87: and if you aim to use the touchscreen with non-android applications (like tslib) you should take note that the tsc2003.c in htc-linux has some calibration algorythm for android use |
14:24.19 | cr2 | ah. they use VirtualSetAttributes(r0,r1,0xb3,... on DEX |
14:24.33 | dcordes | ex87: http://headfuzz.co.uk/files/android/git-diff-2.6.25-170908.diff here's a patch against android-msm-2.6.25 |
14:24.36 | dcordes | you can use |
14:24.57 | dcordes | then you get 2.6.25 base with htc-linux stuff plus updates |
14:25.03 | cr2 | 0xb3,0x7ff |
14:25.15 | *** join/#htc-linux kiozen (n=oeichler@rgnb-5d87cc6d.pool.einsundeins.de) |
14:25.41 | dcordes | ex87: http://linuxtogo.org/~lgorris/git-diff-2.6.25-170908-polaris.diff this is almost the same patch but adds support for polaris and fixes two errors |
14:25.57 | cr2 | pVirtMem, |
14:25.57 | cr2 | <PROTECTED> |
14:26.46 | ex87 | dcordes: Doesn't polaris patch break compability with kaiser? |
14:27.02 | cr2 | VirtualSetAttributes(pVirtMem,dwBytes,0xb3,0x7ff); |
14:28.37 | dcordes | ex87: no it shouldn't. but you can try. if it doesn't work use git-reset --hard and apply the headfuzz patch |
14:28.39 | cr2 | Kevin2: i'm naming all irqs now, will test it, and pastbin the patch. |
14:29.01 | dcordes | ex87: if you only look for improving the touchscreen, htc-linux is sufficient |
14:29.11 | dcordes | because nothing about it changes |
14:29.44 | dcordes | ex87: do you want the non-android touchscreen patch? |
14:29.57 | Kevin2 | cr2: Great! |
14:31.49 | dcordes | ex87: http://linuxtogo.org/~lgorris/tsc2003-kaiser-tslib.diff |
14:31.56 | ex87 | dcordes: haven't tried to use non-androids, but it could be useful for later days |
14:32.00 | ex87 | thanks |
14:32.47 | dcordes | I'm preparing a new angstrom based image with tslib and tslib test programs |
14:33.18 | dcordes | no problem |
14:33.19 | dcordes | bbl |
14:35.28 | *** join/#htc-linux sikkle (n=sikkle@modemcable132.57-201-24.mc.videotron.ca) |
14:40.54 | NetRipper | once the linux kernel boots, how does that write output to the "terminal"? also using raw framebuffer? |
14:41.18 | NetRipper | and for it to work, does it rely on haret initializing framebuffer properly? |
14:41.31 | NetRipper | cr2 or Kevin2, any of you know? |
14:42.11 | NetRipper | on my hermes it's just haret output and then right below it is kernel output, doesn't seem like a special kind of initialization by linux kernel |
14:42.14 | cr2 | NetRipper: with vfsb - yes |
14:42.39 | cr2 | NetRipper: but it's better to have a normal video driver, otherwise you can't suspend/resume ;-) |
14:43.10 | NetRipper | googles vfsb |
14:43.14 | *** join/#htc-linux dcordes_ (n=dcordes@unaffiliated/dcordes) |
14:44.37 | cr2 | NetRipper: vsfb |
14:44.46 | NetRipper | cr2, basically, if linux were to boot, even though framebuffer code of haret doesn't display, should we start seeing linux output? |
14:44.47 | cr2 | VerySimpleFrameBuffer |
14:44.50 | NetRipper | ah |
14:45.10 | cr2 | that's not likely to happen |
14:45.21 | cr2 | we have msm_fb from android |
14:45.37 | cr2 | but we need a specific mddi panel init |
14:45.41 | cr2 | it's missing there. |
14:45.45 | NetRipper | hm |
14:45.54 | cr2 | there may be 3 panels on raph100 |
14:46.05 | cr2 | hitachi, toppoly and sharp |
14:46.17 | cr2 | toppoly is (as always) the biggest pita |
14:46.26 | NetRipper | mddi is described on kaiser memory map |
14:46.43 | cr2 | it needs more setup |
14:47.03 | cr2 | mddi for kaiser is too limited, we need a better description |
14:47.26 | cr2 | toshiba-android is also a very incomplete setup. |
14:52.32 | swetland | the mddi stuff is a total pita |
14:52.53 | swetland | you have to talk to the mddi controller, to issue commands to the bridge chip, which then often uses gpio, spi, etc to talk to the panel |
14:54.43 | cr2 | swetland: or i2c ;-) |
14:55.07 | swetland | all kinds of fun |
14:55.11 | cr2 | and it's all panel-dependent |
14:55.18 | swetland | hate lcd panel init. it is always incredibly hateful |
14:55.36 | swetland | can't think of any hardware I've ever dealt with where it didn't suck |
14:56.10 | NetRipper | yea why can't you just set a bit and it'll turn on and set default values by itself ;) |
14:56.14 | cr2 | :( |
15:01.28 | cr2 | Kevin2: addlist gpios syntax includes masking ? |
15:01.45 | cr2 | Kevin2: it seems i've mixed it with irqs |
15:02.39 | cr2 | ah, yes. |
15:04.12 | Kevin2 | The GPIOS and IRQS use the same variable types. |
15:05.00 | Kevin2 | (So, you can do the same thing with IRQS - run a series of "ibit IRQS ..." and then run "show IRQS" to see the mask.) |
15:07.38 | cr2 | Kevin2: but the addlist syntax is the same ? |
15:09.30 | cr2 | 006.145 IRQS IRQ: MDP(19)=1 |
15:09.30 | cr2 | 006.155 IRQS IRQ: MDDI_PRI(16)=1 |
15:09.34 | cr2 | this looks good |
15:09.50 | cr2 | but something else is broken |
15:10.28 | cr2 | IRQ: ?(33)=2 |
15:14.57 | cr2 | Kevin2: show irqs |
15:15.00 | cr2 | HaRET(20)# show irqs |
15:15.00 | cr2 | <PROTECTED> |
15:15.00 | cr2 | <PROTECTED> |
15:15.00 | cr2 | <PROTECTED> |
15:15.00 | cr2 | <PROTECTED> |
15:16.21 | cr2 | sound goes over ADSP_A11 |
15:16.46 | cr2 | but the FM radio uses ADM_AARM |
15:16.51 | cr2 | and ?52 |
15:17.06 | Kevin2 | cr2: What's your .py stuff look like? |
15:19.26 | cr2 | i'll pastebin |
15:20.56 | cr2 | http://pastebin.ca/1212836 |
15:22.01 | Kevin2 | cr2: irqs1 bit numbers need to be between 0-31. (It's the bitoffset of that register.) |
15:22.20 | cr2 | ok |
15:23.53 | Kevin2 | BTW, you don't want to give two registers the same name (ie, you want IRQ0/IRQ1 not IRQ/IRQ). The mmutrace stuff will only show the irq name, and if you use the same name for two registers, you wont be able to tell which one got accessed. |
15:24.27 | cr2 | wee :) |
15:24.28 | cr2 | ok |
15:24.29 | Kevin2 | s/irq name/register name/ |
15:26.18 | cr2 | yes, i've corrected it. |
15:26.27 | *** join/#htc-linux drasar (n=maik@77.93.211.13-msol.msol.cz) |
15:27.04 | cr2 | 000.000 IRQS IRQ0: ADM_AARM(21)=1 |
15:27.04 | cr2 | 000.000 IRQS IRQ1: UART2DM_RX(52)=1 |
15:27.18 | cr2 | but i dont see the gpios |
15:27.25 | cr2 | gpio irqs |
15:27.42 | cr2 | Kevin2: what does 'last' mean ? |
15:27.45 | cr2 | <PROTECTED> |
15:27.45 | cr2 | <PROTECTED> |
15:28.23 | Kevin2 | "last" means only report on change. You want "0" there for IRQS (report on any non zero value). |
15:29.07 | Kevin2 | cr2: The high order bits of the gpio registers are used, and the low order bits are bogus? |
15:29.36 | cr2 | no :) |
15:29.44 | cr2 | should i invert them ? |
15:29.57 | Kevin2 | Yes - the mask lists those bits to be ignored. |
15:30.25 | Kevin2 | BTW, why do you need to mask the registers? Will the high bits ever be non zero? |
15:31.36 | cr2 | don't know, but they are not used |
15:32.53 | Kevin2 | It probably isn't needed to mask them then. Wont hurt either. |
15:33.56 | Kevin2 | Hrmm. Do they really call the first gpio in bank 2 - "gpio43"? |
15:34.47 | cr2 | yes, it's the running number |
15:35.03 | cr2 | http://wiki.xda-developers.com/index.php?pagename=Raphael_GPIO |
15:39.36 | drasar | Hello. Can somebody tell me why vanilla kernel doesn't boot on himalaya (or why isn't himalaya specific code ported to mainline)? |
15:39.48 | cr2 | Kevin2: got that now -> |
15:39.51 | cr2 | 001.934 IRQS irq0: GPIO0(64)=1 GPIO1(65)=1 GPIO2(66)=1 GPIO3(67)=1 GPIO4(68)=1 GPIO5(69)=1 GPIO6(70)=1 GPIO7(71)=1 GPIO8(72)=1 GPIO9(73)=1 GPIO10(74)=1 GPIO11(75)=1 GPIO12(76)=1 GPIO13(77)=1 GPIO14(78)=1 GPIO15(79)=1 |
15:41.03 | cr2 | arch/arm/mach-msm/gpio_hw.h says 1=interrupt pending |
15:42.26 | cr2 | HaRET(8)# vd 0xb2f008a0 4 |
15:42.26 | cr2 | b2f008a0 | 0000ffff | .... |
15:42.26 | cr2 | HaRET(9)# pd 0xa92008a0 4 |
15:42.26 | cr2 | a92008a0 | 0000ffff | .... |
15:42.26 | cr2 | HaRET(10)# |
15:42.32 | cr2 | ?? |
15:43.50 | cr2 | drasar: -hh kernel should boot on hima, but many other drivers are missing. |
15:48.09 | *** join/#htc-linux sikkle (n=sikkle@modemcable132.57-201-24.mc.videotron.ca) |
15:50.48 | drasar | cr2: Yes, I know, but is any reason why isn't code from hh kernel ported to vanilla? |
15:51.53 | cr2 | asic3 |
15:52.21 | Kevin2 | cr2: Hrmm. Maybe the values are only valid when an irq is actually pending? |
15:52.40 | cr2 | Kevin2: i can mmutrace that register |
15:53.23 | cr2 | watch gpios shows the 'vsync' gpio |
15:53.25 | cr2 | 004.206 GPIOS in4: in4-2(290)=1 |
15:53.30 | cr2 | it's address is |
15:55.12 | Kevin2 | Those addresses aren't actually hardware addresses. It looks to be just shared memory to the arm9.. |
15:56.02 | Kevin2 | If you mmutrace 0xb2f008a0 do you see reads of something other than 0xffff ? |
15:56.15 | cr2 | i got an exception ? |
15:56.25 | cr2 | hm. ok |
15:56.45 | cr2 | HaRET(20)# vd 0xb2f008ac 4 |
15:56.45 | cr2 | b2f008ac | 000000fc | .... |
15:57.36 | cr2 | HaRET(23)# show mmutrace |
15:57.36 | cr2 | 000: 0xb2f008ac 4 'rw' 00000000 none |
15:58.07 | cr2 | now i need to purge 2: 0xb2f008a0 ffff0000 32 0 |
15:58.14 | cr2 | is it possible ? |
15:58.34 | Kevin2 | Huh? |
15:58.52 | cr2 | i've added it to irqs list |
15:58.56 | cr2 | before |
15:59.07 | Kevin2 | I just use clearvar IRQS and start over. I don't remember if there was a way to clear just one item. |
15:59.30 | cr2 | clear irqs 2 ? |
15:59.55 | cr2 | no, it removed everything |
16:01.04 | cr2 | btw, 0: 0x95300080 00000100 32 0 |
16:01.16 | cr2 | is 00000100 something hermes-specific ? |
16:01.22 | cr2 | s/hermes/kaiser/ |
16:02.05 | Kevin2 | cr2: It's the timer irq on vogue I think. I think you said it was the timer irq on your device also. I usually mask the timer, because it just triggers way too often. |
16:03.28 | cr2 | ok, need to check it. |
16:03.30 | cr2 | 009.514 IRQS IRQ1: GPIO2(33)=1 |
16:03.59 | cr2 | don't see the changes of gpio status |
16:06.17 | Kevin2 | How about doing a: addlist TRACES 0xb2f008ac |
16:06.24 | Kevin2 | And then running wirq |
16:07.56 | cr2 | Kevin2: watch gpios sees it, but mmutrace not. |
16:07.59 | cr2 | trying |
16:08.55 | cr2 | doesn't work too |
16:09.12 | Kevin2 | Hrmm. Isn't 8ac bank 4? |
16:09.37 | cr2 | 002.692 GPIOS in4: in4-2(290)=0 |
16:09.37 | cr2 | 002.694 GPIOS in4: in4-2(290)=1 |
16:09.37 | cr2 | 003.014 GPIOS in4: in4-2(290)=0 |
16:09.37 | cr2 | 003.016 GPIOS in4: in4-2(290)=1 |
16:10.06 | Kevin2 | That doesn't mean they generated an irq. |
16:10.17 | cr2 | ok, that#s true. |
16:10.54 | cr2 | no irqs in bank4 |
16:10.56 | cr2 | http://wiki.xda-developers.com/index.php?pagename=Raphael_GPIO |
16:11.03 | cr2 | need to pick some other |
16:11.54 | cr2 | bank3 gpio94 in3-26 |
16:12.17 | cr2 | 001.916 GPIOS in3: in3-26(282)=1 |
16:12.17 | cr2 | 001.924 GPIOS in3: in3-26(282)=0 |
16:12.39 | cr2 | +0x8a8 |
16:13.49 | cr2 | 003.683 IRQS IRQ1: GPIO2(33)=1 |
16:13.49 | cr2 | 003.696 IRQS IRQ1: GPIO1(32)=1 |
16:13.49 | cr2 | 003.696 8006aff8: e5922000(ldr) # irq3==078f400f |
16:13.50 | cr2 | 003.759 IRQS IRQ1: GPIO1(32)=1 |
16:15.44 | Kevin2 | What did you define "irq3" as? |
16:15.48 | cr2 | too much irqs without maskink |
16:16.12 | cr2 | bank3 irq stat |
16:16.49 | cr2 | need ~0x7ffffff now |
16:17.00 | cr2 | 0x7ff ffff |
16:17.15 | cr2 | 0x07ff ffff |
16:17.24 | Kevin2 | What does that register look like if you just read it with pd/vd? |
16:17.30 | cr2 | 0xfff800 0000 |
16:18.12 | cr2 | HaRET(58)# vd 0xb2f008a8 4 |
16:18.12 | cr2 | b2f008a8 | 038f400f | .@.. |
16:20.04 | cr2 | http://git.android.com/?p=kernel.git;a=blob;f=arch/arm/mach-msm/gpio.c;h=93bf7264b487bbeb9c9d9a1c000949f0076d0e57;hb=android-msm |
16:20.58 | Kevin2 | cr2: What does "vd 0xb2f00888 4" (in_en) report? |
16:21.10 | Kevin2 | s/in_en/int_en/ |
16:21.54 | cr2 | HaRET(59)# vd 0xb2f00888 4 |
16:21.54 | cr2 | b2f00888 | 04408000 | ..@. |
16:22.37 | Kevin2 | What about 0xb2f0088c? |
16:22.52 | cr2 | HaRET(60)# vd 0xb2f0088c 4 |
16:22.53 | cr2 | b2f0088c | 00000000 | .... |
16:23.46 | cr2 | <PROTECTED> |
16:23.46 | cr2 | <PROTECTED> |
16:23.46 | cr2 | <PROTECTED> |
16:23.46 | cr2 | <PROTECTED> |
16:23.48 | cr2 | ? |
16:24.19 | Kevin2 | Ughh. It looks like interrupt status is only valid when and'ing with the interrupt enable. |
16:33.13 | cr2 | ok, but still moved a bit forward |
16:37.07 | *** join/#htc-linux exco (n=exco@e181072093.adsl.alicedsl.de) |
17:20.53 | *** join/#htc-linux LunohoD_ (n=alex@e180078171.adsl.alicedsl.de) |
17:44.05 | *** join/#htc-linux exco (n=exco@e181072093.adsl.alicedsl.de) |
17:44.34 | *** join/#htc-linux linkyboy (n=a@c-75-73-235-231.hsd1.mn.comcast.net) |
17:44.45 | linkyboy | hello, will running android in my Kaiser hard reset it? |
17:50.19 | exco | you can run into problems forcing you to hard reset |
17:50.43 | exco | at least that's what I figure from reading the posts on xda-developers |
17:55.59 | cr2 | exco: hard reset == factory restore ? |
18:05.38 | exco | think so |
18:05.48 | exco | but I guess he already left, cr2 |
18:06.12 | exco | btw do you know if ph5 already built an fso image for magician |
18:11.00 | *** join/#htc-linux LunohoD_ (n=alex@e180071059.adsl.alicedsl.de) |
18:13.32 | cr2 | ~seen ph5 |
18:13.35 | apt | ph5 <n=ph5@p5485EEE9.dip.t-dialin.net> was last seen on IRC in channel #maemo, 29d 5h 15m 32s ago, saying: 'hi, does anybody know if there exists a tango style device icon depicting a nokia n810 somewhere?'. |
18:14.02 | cr2 | didn't hear from him since july |
18:33.21 | dcordes_ | hi |
18:34.47 | dcordes_ | ex87: everything went well with the kernel compilation? |
18:42.36 | cr2 | dcordes_: can you remind me how do we document the SMD channels ? |
18:43.07 | cr2 | +0x041f0 8k+24 AT commands send channel |
18:43.43 | cr2 | it's the start ? usuall 0x20000000 |
18:50.45 | *** join/#htc-linux timebomb (n=timebomb@e176104187.adsl.alicedsl.de) |
18:53.30 | *** join/#htc-linux lpotter (n=ljp@CPE-124-191-144-181.vic.bigpond.net.au) |
18:56.41 | BabelO | kiozen: wouah :) want a patch |
18:56.53 | BabelO | kiozen: that's what we see last time :) |
18:57.37 | BabelO | kiozen: btw , i have small problem ith line detection, trying another algo. it seem your map give me to much noise |
18:57.49 | kiozen | BabelO: we can make a bet who files a patch first :) |
18:58.12 | kiozen | BabelO: the scanned one or the official one? |
18:59.11 | BabelO | kiozen: the scanned , my scanned works well because color are good, on your scanned map line are truncated and color is not strong |
18:59.55 | BabelO | kiozen: bt i ve written a class for image manipulation now, filtering, projection and some other stuff like auto threshold |
18:59.58 | kiozen | BabelO: that's what I expected :) also you can see the print raster as it is scanned with 600dpi, tought map |
19:00.33 | kiozen | BabelO: sounds cool :) looking forward to your patches! |
19:00.54 | kiozen | BabelO: have done some bug fixing in GT, track related stuff. |
19:01.48 | kiozen | adms family is waiting ;) bbl |
19:07.14 | *** join/#htc-linux exco (n=exco@e181072093.adsl.alicedsl.de) |
19:10.27 | cr2 | 0x18 or 0x19 |
19:10.42 | cr2 | *2= 0x30... |
19:10.55 | cr2 | or 0x32 - unlikely |
19:12.04 | dcordes_ | cr2: yes, 0x041f0 is where the first AT data shows |
19:23.41 | cr2 | reset is ok. |
19:24.48 | cr2 | the accelerometer datasheet does not match the htc driver 100% ;) |
19:27.20 | dcordes_ | took a close look on it? |
19:28.08 | cr2 | yes. |
19:29.43 | cr2 | it uses some not documented regs |
19:36.32 | cr2 | NetRipper: can you write a test program using accelerometer wince api for haret ? |
19:36.54 | cr2 | dcordes_: still didn't find the digital compass... |
19:39.34 | cr2 | dcordes_: the referenced datasheet is very useful, but an actual device is a bit different. |
19:40.37 | *** join/#htc-linux Dinde36 (n=kayser@81-65-130-197.rev.numericable.fr) |
19:44.05 | BabelO | je vais m asperger de sel |
19:44.12 | BabelO | :) |
19:44.18 | BabelO | wrong windows again |
19:46.43 | NetRipper | cr2, native c/cpp or can it be .net? |
19:49.20 | cr2 | NetRipper: haret is .cpp, but i guess it'll be native .c |
19:50.07 | NetRipper | cr2, oh you want it within the haret.exe? |
19:50.12 | cr2 | but i've looked at the driver already, there is not much that one can trace. only to confirm the i2c traffic |
19:50.57 | cr2 | write something, we will merge it into haret (GSENSTST or something like that) |
19:51.10 | cr2 | later |
19:53.26 | cr2 | i'm looking at NAVI aka pSOC aka smarttouch aka microP aka cap_sens |
19:56.33 | NetRipper | cr2, a simple app using the htcsensorsdk.dll to query gsensor data, shouldn't be too hard |
19:57.24 | cr2 | ok |
19:57.48 | cr2 | do you have any idea about digital compass ? |
20:00.38 | cr2 | KLT1: is this navi |
20:02.04 | cr2 | hehe. reset psoc by microp |
20:02.25 | NetRipper | dont know about navi or compass |
20:07.01 | cr2 | this navi driver looks big an complex. |
20:09.50 | cr2 | i gave up for now :) |
20:11.25 | cr2 | NetRipper: sensorsdk can control the light sensor too ? |
20:11.26 | NetRipper | hehe |
20:11.28 | NetRipper | yes |
20:12.44 | NetRipper | in c# i can make a test app real quick, in c/cpp i'll need some more time |
20:16.27 | cr2 | ? |
20:16.41 | NetRipper | for gsensor |
20:16.41 | cr2 | you need to call the c api in # anyway |
20:16.56 | NetRipper | in #? |
20:17.04 | cr2 | the gui may be in c#, who cares |
20:17.29 | NetRipper | indeed |
20:17.30 | NetRipper | k |
20:20.16 | cr2 | lol |
20:20.18 | *** join/#htc-linux Dinde (n=kayser@81-65-130-197.rev.numericable.fr) |
20:20.30 | cr2 | StylusOutStatus |
20:21.46 | *** join/#htc-linux yesher (n=yeshtar@98.109.131.116) |
20:23.29 | cr2 | SPM1: for keybd |
20:31.39 | ALoGeNo | hi there |
20:41.46 | *** join/#htc-linux tgreer (n=tgreer@titus.tsone.info) |
20:42.49 | *** join/#htc-linux pleemans (n=toi@d54C2AAB7.access.telenet.be) |
20:46.24 | *** join/#htc-linux julien_ (n=julien@157.208.20.81.dynamic.adsl.abo.nordnet.fr) |
20:52.40 | *** join/#htc-linux miknix (n=miknix@bl5-240-155.dsl.telepac.pt) |
20:57.27 | cr2 | the microP will be an interesting puzzle |
20:57.42 | cr2 | the big keyboard is served over i2c |
20:58.27 | Espenfjo | Wonder if i should take up work on the diamond again |
20:58.54 | cr2 | sure |
20:59.27 | Espenfjo | Just got so demotivated when i couldnt get the framebuffer from haret updating correctly :) |
20:59.30 | cr2 | Espenfjo: you don't have this big keyboard (and SD), but the rest should be the same. |
20:59.48 | cr2 | it's a minor problem |
21:00.02 | Espenfjo | hehe, well.. :) dunno how many hours i have used on making it work this far |
21:00.06 | cr2 | the big problem is to teach msm_fb to initialize the panel. |
21:00.28 | cr2 | Espenfjo: can you do a soft reset |
21:00.31 | cr2 | and then |
21:00.54 | cr2 | pwf diam-dmesg 0x17200000 0x100000 |
21:01.08 | *** join/#htc-linux marmotta (n=skodde@unaffiliated/skodde) |
21:01.16 | Espenfjo | yeah, sure, hold on |
21:01.52 | cr2 | i also do not understand why there are SMEM layout differences between raphael and diamond |
21:02.33 | Espenfjo | yeah, it seems like there are some weird differences |
21:05.22 | Espenfjo | pl. did the pwf now |
21:05.23 | Espenfjo | ok |
21:05.44 | cr2 | look at it with the text editor |
21:07.58 | Espenfjo | yeah |
21:09.01 | Espenfjo | What should i look for? |
21:10.00 | cr2 | lcd panel type |
21:11.13 | NetRipper | cr2, http://www.netripper.nl/raphael/SensorTestDeploy.CAB |
21:13.29 | Espenfjo | The only thing which seem remotly like a panel type is: 23:02:00 [D:DISP] g_CurrentPanelType=10... |
21:14.25 | cr2 | 10 ? |
21:14.53 | Espenfjo | yeah |
21:15.19 | *** join/#htc-linux dcordes (n=dcordes@e179106165.adsl.alicedsl.de) |
21:16.16 | Espenfjo | the next line mentions toshiba_prim start, although not sure if it has any relevance to the panel |
21:17.55 | Espenfjo | also; [HDS_Init] Not Hitachi panel. |
21:18.00 | Espenfjo | But thats all |
21:18.58 | *** join/#htc-linux miknix (n=miknix@bl5-240-155.dsl.telepac.pt) |
21:19.49 | cr2 | NetRipper: very nice :) the light sensore is in the speaker hole |
21:19.58 | cr2 | Espenfjo: toppoly ? sharp ? |
21:20.13 | Espenfjo | dunno, it doesnt say anything |
21:20.28 | cr2 | grep -i for these strings |
21:20.40 | *** part/#htc-linux tgreer (n=tgreer@titus.tsone.info) |
21:20.43 | Kevin2 | cr2: I think I'm going to tag the current haret as v0.5.2. |
21:20.44 | NetRipper | yep that's the light sensor |
21:20.59 | Espenfjo | yeah, already done ofc. :) but no, doesnt get anthing |
21:21.30 | cr2 | NetRipper: the data is a bit strange |
21:21.42 | cr2 | Kevin2: .py irq names for msm ? |
21:21.53 | NetRipper | cr2, how is it strange? |
21:22.11 | NetRipper | besides the orientation, which doesn't seem to change |
21:22.23 | Kevin2 | cr2: Do you want me to put them in before 0.5.2? I'd just assume add them afterwards. |
21:22.31 | cr2 | NetRipper: the data is noisy. sstrange is a bad word |
21:23.02 | NetRipper | it's the raw sensor data |
21:23.05 | cr2 | Kevin2: i don't see anything speaking against including it. the gpio irq is another issue |
21:23.22 | NetRipper | and the 0 flipping to 359 is normal |
21:23.36 | NetRipper | being degrees |
21:23.37 | NetRipper | and all |
21:23.41 | Kevin2 | cr2: Do you have a patch ready? |
21:23.49 | cr2 | NetRipper: great. we will have something to compare |
21:24.18 | cr2 | Kevin2: i'll update from CVS, and will make a diff |
21:27.45 | cr2 | Kevin2: http://pastebin.ca/1213076 |
21:29.43 | Kevin2 | cr2: Okay. Why comment out the gpio names? I don't think it will hurt if haretconsole knows the names. |
21:29.51 | cr2 | ok. |
21:31.31 | cr2 | Kevin2: i think we should think about haret-1.0, and setting the goals and tasks. |
21:33.10 | Kevin2 | I think Haret will always be a developer tool - as such, I think we can call it 1.0 whenever we want. :-) |
21:33.22 | cr2 | lol |
21:43.14 | Kevin2 | Okay - I tagged 0.5.2 - now to update the links. |
21:43.38 | ALoGeNo | hey to do the kernel boot in rootfs.img i need set up ramfs in config no? |
21:43.43 | ALoGeNo | and lo all :) |
21:43.57 | ALoGeNo | in the uni i mean |
21:44.42 | cr2 | Kevin2: nice. |
21:44.55 | cr2 | Espenfjo: i'll look at the diamond |
21:45.18 | Espenfjo | :) |
21:46.29 | cr2 | aah. this i2c panel setup goes to microP |
21:46.38 | cr2 | cc,20,c |
21:47.08 | cr2 | cc and ce. maybe 2 microp. atmega48 and atmega88 |
21:49.03 | *** join/#htc-linux nebi (n=nebi@c-498fe253.02-145-7570701.cust.bredbandsbolaget.se) |
21:51.19 | cr2 | Espenfjo: hitachi/sharp/toppoly. looks like the same setup as on raph100 |
21:51.45 | Espenfjo | ok |
21:52.09 | cr2 | the same microp48 and microp88 |
21:52.19 | Espenfjo | ok, that sounds promising |
21:52.47 | cr2 | Espenfjo: you should look here too -> http://wiki.xda-developers.com/index.php?pagename=Raphael_GPIO |
21:53.04 | cr2 | and trace the gpios. |
21:53.19 | cr2 | let's hope they are the same. |
21:54.44 | cr2 | Espenfjo: try to suspend and resume. then dump 0x172* again. |
21:54.48 | Espenfjo | I could try, tomorrow maybe. But its a bit out of my league :) I have no idea what to do. I know my C, and have some basic HW understanding and i would say i know fairly much about the linux kernel also |
21:54.59 | cr2 | let's hope you will see the LCD type there. |
21:55.08 | Espenfjo | tomorrow :) have to get some sleep before work |
21:55.18 | cr2 | Espenfjo: ok. good night :) |
21:56.34 | cr2 | it does the same TSSC ops, so the touchscreen should be the same. |
21:57.30 | cr2 | hehe. maybe we don't really need 2 machtypes here :) |
21:57.50 | cr2 | it's like hermes and trinity |
22:04.21 | *** join/#htc-linux patoche (n=pat@lns-bzn-20-82-64-10-164.adsl.proxad.net) |
22:04.29 | *** part/#htc-linux patoche (n=pat@lns-bzn-20-82-64-10-164.adsl.proxad.net) |
22:07.20 | cr2 | lol, i've read about the possibility for G1 to run wince. |
22:09.42 | yesher | have there been any ce devices with compass and accelerometer before? |
22:09.49 | BabelO | cr2: windoze guy are crazy :) |
22:10.01 | BabelO | cr2: they want to add more bug |
22:11.49 | dcordes | cr2: what about that? there are also people who pay money for wince on the beagleboard |
22:14.29 | dcordes | cr2: do you have details about G1 boot process? |
22:23.26 | Marajin | yoes |
22:23.33 | NetRipper | #htc-wm? :p |
22:30.30 | cr2 | BabelO: i wish them good luck ;-) it's even a bigger pita than porting linux. |
22:30.53 | BabelO | cr2: yes i think so |
22:30.53 | cr2 | Marajin: do you know any wince device with a digital compass ? |
22:31.13 | Marajin | cr2: nope |
22:31.17 | cr2 | i'm confused by the compass irq, and don't know how to trigger it :) |
22:31.44 | cr2 | the accelerometer will work. |
22:32.05 | BabelO | cr2: do you want to participate to that -> http://www.project10tothe100.com/intl/EN_GB/index.html |
22:32.13 | BabelO | cr2: linux for the whole world :) |
22:32.13 | cr2 | datasheet is not very accurate (maybe it's another device), but it's more of less clear how to read the data |
22:33.54 | cr2 | BabelO: osm is certainly a big concern for google. since they don't allow to draw using their imagery ;) |
22:34.48 | BabelO | cr2: really ? but i see some comment about VirtualEarth and a third |
23:03.58 | BabelO | good night |
23:23.39 | *** join/#htc-linux sikkle (n=sikkle@modemcable132.57-201-24.mc.videotron.ca) |
23:36.47 | *** join/#htc-linux sikkle (n=sikkle@modemcable132.57-201-24.mc.videotron.ca) |
23:38.36 | *** join/#htc-linux ALoGeNo (n=QUAKEIII@26.Red-81-37-222.dynamicIP.rima-tde.net) |